International Journal of Obesity (2013) 37, 216 - 223 & 2013 Macmillan Publishers Limited All rights reserved 0307-0565



The type and quantity of dietary fat and carbohydrate alter faecal microbiome and short-chain fatty acid excretion in a metabolic syndrome ‘at-risk’ population
F Fava1,2, R Gitau1, BA Griffin3, GR Gibson1, KM Tuohy1,2, and JA Lovegrove1 INTRODUCTION AND OBJECTIVES: An obese-type human microbiota with an increased Firmicutes:Bacteroidetes ratio has been described that may link the gut microbiome with obesity and metabolic syndrome (MetS) development. Dietary fat and carbohydrate are modifiable risk factors that may impact on MetS by altering the human microbiome composition. We determined the effect of the amount and type of dietary fat and carbohydrate on faecal bacteria and short chain fatty acid (SCFA) concentrations in people ‘at risk’ of MetS. DESIGN: A total of 88 subjects at increased MetS risk were fed a high saturated fat diet (HS) for 4 weeks (baseline), then randomised onto one of the five experimental diets for 24 weeks: HS; high monounsaturated fat (MUFA)/high glycemic index (GI) (HM/HGI); high MUFA/low GI (HM/LGI); high carbohydrate (CHO)/high GI (HC/HGI); and high CHO/low GI (HC/LGI). Dietary intakes, MetS biomarkers, faecal bacteriology and SCFA concentrations were monitored. RESULTS: High MUFA diets did not affect individual bacterial population numbers but reduced total bacteria and plasma total and LDL-cholesterol. The low fat, HC diets increased faecal Bifidobacterium (P ¼ 0.005, for HC/HGI; P ¼ 0.052, for HC/LGI) and reduced fasting glucose and cholesterol compared to baseline. HC/HGI also increased faecal Bacteroides (P ¼ 0.038), whereas HC/LGI and HS increased Faecalibacterium prausnitzii (P ¼ 0.022 for HC/HGI and P ¼ 0.018, for HS). Importantly, changes in faecal Bacteroides numbers correlated inversely with body weight (r ¼ À0.64). A total bacteria reduction was observed for high fat diets HM/HGI and HM/LGI (P ¼ 0.023 and P ¼ 0.005, respectively) and HS increased faecal SCFA concentrations (Po0.01). CONCLUSION: This study provides new evidence from a large-scale dietary intervention study that HC diets, irrespective of GI, can modulate human faecal saccharolytic bacteria, including bacteroides and bifidobacteria. Conversely, high fat diets reduced bacterial numbers, and in the HS diet, increased excretion of SCFA, which may suggest a compensatory mechanism to eliminate excess dietary energy. International Journal of Obesity (2013) 37, 216 -- 223; doi:10.1038/ijo.2012.33; published online 13 March 2012 Keywords: RISCK; gut microbiota; dietary fat; dietary carbohydrate; SCFA

INTRODUCTION The prevalence of both obesity and metabolic syndrome (MetS) is rising exponentially on a global scale.1 MetS is a constellation of characteristics (dyslipidaemia: elevated plasma triacylglycerol (TAG) and low HDL-cholesterol levels, hypertension, central adiposity, insulin resistance),2 - 4 which confers increased coronary heart disease risk. Diet has an important role in MetS pathogenesis. There is evidence that the amount and type of dietary fats and carbohydrates (CHO) can modify dyslipidaemia, insulin sensitivity, endothelial dysfunction and blood pressure.5,6 Recently, the human gut microbiota has been implicated in MetS risk.7,8 The gut microbiota appears to differ between lean and obese animals, whether obesity is diet induced9 - 11 or genetic.12 - 14 Obese-type gut microbiota, characterised by a higher Firmicutes:Bacteroidetes ratio in obese as opposed to lean, have been observed both in murine models of obesity and humans. Moreover, germ-free mice colonised with the microbiota from obese mice display increased body fat, higher faecal total

energy content (by bomb calorimetry) and higher concentrations of faecal short chain fatty acids (SCFA) compared with their conventionally fed lean counterparts, indicating that the microbiota of obese animals may have an increased capacity to harvest energy.15 In addition, weight loss in humans induced by CHOor fat-restricted diets has been associated with a change in gut microbial composition, resembling the microbiota of lean individuals (that is, increased Bacteroidetes).16 The colonic fermentation of carbohydrate and fibre produces SCFA, which may impact on a number of physiological processes related to human energy metabolism, including satiety, hepatic lipogenesis, adipocyte fat deposition and thermogenesis.7 Although studies have examined the impact of selected fibres and prebiotics on the human gut microbiota, few studies have examined the impact of whole dietary carbohydrate load on intestinal bacteria, especially in individuals at MetS risk.17,18 Data from human studies show higher faecal SCFA concentrations in overweight and obese humans compared with their lean

1 Department of Food and Nutritional Sciences and Institute for Cardiovascular and Metabolic Research (ICMR), The University of Reading, Whiteknights, Reading, UK; 2IASMA Research and Innovation Centre, Fondazione Edmund Mach, Department of Food Quality and Nutrition, Nutrition and Nutrigenomics Unit, San Michele all’Adige, Italy and 3 Faculty of Health and Medical Sciences, University of Surrey, Guildford, UK. Correspondence: Dr F Fava, IASMA Research and Innovation Centre, Fondazione Edmund Mach, Department of Food Quality and Nutrition, Nutrition and Nutrigenomics Unit, Via E. Mach n. 1, 38010 San Michele all’Adige, Trento, Italy. E-mail: Received 7 December 2011; revised 24 January 2012; accepted 29 January 2012; published online 13 March 2012

diameter ¼ 0. participants were scored on their BMI or waist circumference. The study was conducted according to the Declaration of Helsinki guidelines and was given a favourable ethical opinion by the University of Reading Research Ethics Committee and the Local Research Ethics Committee. HC/LGI:high CHO/ low GI) for 24 weeks.53 mm. This might suggest the existence of some compensatory mechanism to prevent weight gain or uncontrolled energy intake via the excretion of excess energy in the form of end-products of fermentation. specific for all Atopobium. University of Reading.31 . A weighted scoring system.20 .2) at 4 1C overnight. MUFA 20%E. Eubacterium rectale/Clostridium coccoides population and Bifidobacterium spp.223 . Standard solutions containing 20. specific for most lactobacilli. 5.. In brief. controlled. HM/LGI: total fat 38%E. specific for Desulfovibrioaceae. PUFA 6%E. then homogenized in a stomacher (Seward STOMACHER 400 Lab System. Supernatants containing bacterial cells were fixed in 4% (wt/vol) paraformaldehyde solution (pH 7. plasma TAG and HDL cholesterol concentrations and hypertension. All groups were matched for age (mean±s. Thetford. This may have occurred by a mechanism involving the increased production.26. samples were vortexed and centrifuged at 100 g for 2. Cryptobacterium. HC/LGI: total fat 28%E. Eggerthella and Olsenella. MUFA 20%E.22.24 especially in individuals with increased obesity and type 2 diabetes risk. Vagococcus.9 kg mÀ2) and HDL cholesterol concentration (1.23 Although very little is known about the impact of high fat diets on the gut microbiota. and Oridobacter splanchnicus. PUFA 6%E.5 mM external standards and 2 mM of internal standard (2-ethylbutyirc acid) were used. GI 64%. UK) for 2 min at normal speed.28. In addition. Dietary intervention The RISCK food exchange model was developed to achieve the dietary targets of the RISCK intervention that has been reported in detail previously. specific for most bacteria within clostridial cluster I and all members of clostridial cluster II. Fatty acids were determined by gas-liquid chromatography on a Hewlett Packard (Agilent) 5890 Series II GC system (HP. parallel design and was carried out at the Hugh Sinclair Unit of Human Nutrition.18. Collinsella. SFA 10%E. Bif164. Agilent International Journal of Obesity (2013) 216 . GI 64%. SFA 10%E. UK) were employed: EREC482.microbiota interactions in obesity F Fava et al counterparts on a similar Western-style diet. Melisococcus. West Sussex.4 mmol lÀ1). UK) fitted with a FFAP column (30 m  0.25 The study inclusion criteria included men and women. saturated fatty acids. HM/LGI:high MUFA/low GI. Seward Ltd. Restoration of bifidobacteria levels in high-fat fed mice through dietary supplementation with the prebiotic oligofructose significantly reduced metabolic endotoxaemia and MetS development. Bacteroides spp. SFA 10%E. Dietary calculations were based on the National Diet and Nutrition Survey and National Food Survey habitual intake estimates. Glass beads (5 mm diameter) were added. of endogenous glucagon-like peptide-2. specific for Faecalibacterium prausnitzii. as previously described. few studies have investigated the effect of high-fat diets on the levels of dominant members of the human gut microbiota or SCFA output. Tetragenococcus. Fpra655. Barnesiella spp. SFA 18%E. we tested the hypothesis that the type and quantity of dietary fat and carbohydrate significantly affect the gut microbiota and colonic fermentation in human subjects ‘at risk’ or suffering from the MetS. of which 88 participants completed the 24 weeks dietary intervention period and provided faecal samples. However. Cambridge. Bac303. after which they were randomly assigned to either continue with the reference diet or one of four experimental diets (HM/HGI:high monounsaturated fat (MUFA)/high GI. Participant recruitment Participants were recruited at the Hugh Sinclair Unit of Human Nutrition at the University of Reading on the basis of their increased risk for developing the MetS.25. HM/HGI: total fat 38%E. SFA 10%E. PUFA 6%E. centrifuged at 13 000 g for 5 min and filtered through 0.26 and describes the analysis of faecal microbiology on samples collected at one site (University of Reading). CHO 45%E. polycarbonate syringe filter. It has been suggested that the gut microbiota may metabolise dietary fats (producing diacylglycerols from polyunsaturated fats).30 Nine genus.27 Participants followed a 4-week run-in reference diet that was a high saturated fat diet (HS. washed with phosphate buffered solution. METHODS Study design This study was part of a five-centred intervention study.2. which improved epithelial barrier function and reduced intestinal permeability. The study was a randomised.) and induced insulin resistance. Milton Keynes.19 Dietary supplementation studies with the prebiotics inulin or oligofructose in humans and animals have shown that changes in colonic SCFA are accompanied by reduced plasma lipids. plasma insulin or glucose. PUFA 6%E. with normal hepatic and renal function.8±4. CHO 45%E. BMI (28.high glycemic index (GI) diet (38%E fat).6±0.39 40 -6-diamidino-2-phenylindole (DAPI) was employed for total bacterial cell enumeration. the ‘RISCK’ trial. A total of 130 age and sex matched. Department of Nutrition and Health Research. Target nutrients intakes (expressed as percentage of total energy intake) for the five isocaloric diets are summarised as follows: HS: total fat 38%E.Diet . MUFA 11%E. MUFA 11%E. Norfolk.d.26 and was achieved by replacing all exchangeable fats and CHO in the habitual diets of each participant with products specifically chosen or formulated to provide the required fat/carbohydrate intake and composition in each of the intervention diets. we investigated the impact of any gut microbial changes on metabolic and cardiovascular disease biomarkers. specific for bifidobacteria and Parascardovia denticolens.5 years). gender. Enterobacteriaceae probe D. PUFA 6%E.3 with 6 M HCl. CHO 55%E. aged between 30 and 65 years.22 Cani et al. Paralactobacillus. GI 64%. which ensured that volunteers expressed a minimum of two features of the MetS (score X4). In the present study. 10. J&W Scientific.and group-specific 16S rRNA-targeted and 50 -Cy3-labelled oligonucleotide probes (MGW-Biotech. 54±9. Ato291. high dietary fat intake may increase the quantities of fat and bile acids reaching the colon. and all Enterococcus. Oenococcus and Catelliococcus spp.29 Dietary manipulation has been described in detail by Moore et al. & 2013 Macmillan Publishers Limited Faecal SCFA analysis by gas chromatography Samples were acidified to pH 2 .9 reported that high-fat diets significantly lowered the levels of dominant members of the gut microbiota in mice (that is. MUFA 12%E.25 Faecal samples collection and analysis Faecal samples were collected within 2 h of defaecation. Pediococcus. single blind. re-suspended in phosphate buffered solution and ethanol for storage at À20 1C and used for fluorescent in situ hybridisation as previously described. convert primary bile acids into secondary bile acids and impact on the enterohepatic circulation of bile acids and fat absorption from the small intestine.. SFA). specific for most bacteria within the clostridial cluster XIVa. SRB687.2 mm.. specific for most Bacteroides and Prevotella spp. GI 51%. funded by the UK’s Food Standards Agency. Lab158. 217 Analysis of food diaries The 4-day food diaries (including 1 weekend day) were analysed by the Nutrition Epidemiology Group at Human Nutrition Research. UK. CHO 45%E. diluted 1/10 (wt/vol) with sterile 1 M phosphate buffered solution pH 7. GI 53%. This was achieved by measuring fecal microbial composition and SCFA in participants undergoing a 6-month dietary intervention. CHO 55%E. free-living volunteers were recruited.50 mm. RISCK was registered as a clinical trial (ISRCTN29111298). 1 and 0. via prebiotic fermentation. was employed. and Lactococcus lactis. a low grade systemic inflammation and higher plasma endotoxin concentrations. Leuconostoc and Weissella. Chis150. HC/HGI:high carbohydrate (CHO)/high GI. specific for most Enterobacteriaceae. which is detailed in Jebb et al.5 min. mouse intestinal bacteria. defined as metabolic endotoxaemia. HC/HGI: total fat 28%E. Crawley. particularly in those with hyperlipidaemia and hypercholesterolaemia.

BP. respectively).0522. South Queensferry. with a lower GI in both low GI diets (Po0. The confidence interval was 95%. A small. The only difference observed was a significant decrease in NEFA concentration after intervention with HC/LGI compared to the control HS (P ¼ 0.013) (Table 3). Anthropometric measures.microbiota interactions in obesity F Fava et al 218 Technologies Ltd.017). UK).5±9. The two International Journal of Obesity (2013) 216 -. TAG. Plasma total and LDL cholesterol were significantly decreased compared to baseline in all intervention groups (Po0. Plasma insulin was determined by a specific commercial enzyme-linked immunosorbent assay (ELISA) kit (DAKO Diagnostic Ltd. UK) a flame-ionisation detector and with glass wool inserted in the injection port. respectively. Statistical analysis Normal distribution of data was tested by calculating the skewness and kurtosis and by employing the Kolmogorov . decrease in body fat percentage was observed with HC/HGI compared to baseline (P ¼ 0.002 and P ¼ 0. body mass index.002) and also compared with the other intervention groups (HC/HGI vs control HS: P ¼ 0. However.7 5. Jebb et al.. total energy intake decreased after the intervention with both high CHO diets compared to baseline (P ¼ 0. UK). and a concomitant increase in monounsaturated fat (Po0.7±9.8±0.023) and to HC/HGI (P ¼ 0. total cholesterol. The initial oven temperature was 100 1C. UK). UK). Similarly.Diet . Table 1.01). USA. Alpha Laboratories Ltd. plasma leptin by a commercially available ELISA (Quantikine Human Leptin Kit. The injected sample volume was 1 ml. HDL-cholesterol and NEFA). Paired Student’s t-test was used to analyze changes within each experimental group after treatment compared to baseline.001). Plasma TAG. fluoride/oxalate (glucose) and heparinised plasma (leptin and insulin). high-MUFA diets showed a significantly lower saturated fat intake compared with the reference diet (Po0.5±16. Cat n.9 1.01). Correlation between changes in each parameter were measured using Pearson’s regression analysis. HC/LGI vs HM/HGI: P ¼ 0.5 90. Germany).7 3. BBE 1B). then increased to 250 1C at 50 1C per min and finally held at 250 1C for 2 min. The manipulation of GI was also successful.8 -28.5 min.01). plain serum (LDL-cholesterol concentration and serum C-reactive protein).3) 5.0 98. Fatty acid concentrations were calculated by peak integration using Atlas Lab managing software (Thermo Lab Systems.5±0. glucose.4±5. biochemical characteristics and intravenous glucose tolerance test No significant changes in anthropometric and blood pressure measurements were observed between the five diets at the end of intervention (Table 3). HC/LGI vs HM/LGI P ¼ 0. LDL-cholesterol was isolated from the serum samples by selective precipitation (Randox CH1350. mined by a chromogenic assay (Chromogenix AB. The difference between the values measured after treatment and at baseline was introduced in the ANOVA as a dependent variable.9 3. blood pressure.25 There was no significant difference between men and women recruited to the Reading cohort of the RISCK study.4 2 (4.0 28 (62.4±14. Serum C-reactive protein was determined by ELISA (Wako. Co.7) 10 (23. Crumlin.0069.0087.7±1. Neuss.001). The concentration of NEFA increased after intervention & 2013 Macmillan Publishers Limited RESULTS Participant characteristics The characteristics of the participants at screening are shown in Table 1.223 . Germany) and expressed as mmol gÀ1 of faeces.Smirnov test.010) also accompanied by a slight decrease in body weight (P ¼ 0. Abingdon.014).1±10. Multiple comparisons between the dietary groups showed no significant differences in the measured biochemical parameters (Table 3). Plasma NEFA concentration was measured by a commercially available assay (Wako NEFA C kit. total cholesterol and HDL-cholesterol concentrations were determined using commercially available kits on a Monarch Automatic Analyzer ILab 600 (Instrumentation Laboratories Ltd.9 84. Data that were not normally distributed were Log10 transformed (that is. the two highCHO diets achieved the target values for reduced fat intake (Po0.8 81. A significant increase in waist circumference compared to baseline was observed in the HM/LGI group after treatment (P ¼ 0. and higher GI in both high GI diets (P ¼ 0. Antrim. USA).011. non-starch polysaccharide consumption increased compared to baseline. The total flow was 140 ml minÀ1. Warrington. Sweden). Intravenous glucose tolerance test Intravenous glucose tolerance test was performed as previously described by Jebb et al.9 5.2±1. Statistical significance of differences between diets was analysed by the analysis of variance (ANOVA) followed by post hoc Bonferroni’s multiple comparison test. which was replaced by increased carbohydrate intake (Po0. The two high-CHO diets differed in total starch intake (P ¼ 0. The head pressure was set at 10 psi and the split ratio was 10:1. Minneapolis. MN. Randox Laboratories Ltd. HC/HGI vs HM/HGI P ¼ 0. Injector and detector temperature were set at 280 1C and 300 1C.2 F (n ¼ 45) 59. Pasadena.3±0.3±29. Abbreviations: BMI.02. raised to 150 1C at 8 1C per min. Plasma PAI-1 (plasminogen activator inhibitor-1) and ICAM-1 (intercellular adhesion molecule 1) were deter¨ lndal. bacterial numbers) before parametric statistical analysis.2) 29.0396. The dietary targets of the RISCK study were broadly met.05).2 Characteristics Age (y) Postmenopause (n (%)) BMI (kg mÀ2) Waist circumference (cm) Total cholesterol (mmol lÀ1) LDL-cholesterol (mmol lÀ1) TAG (mmol lÀ1) Glucose (mmol lÀ1) Insulin (pmol lÀ1) Systolic BP (mm Hg) Diastolic BP (mm Hg) Cigarettes smokers (n (%)) BP medication (n (%)) Metabolic score Blood samples collection and analysis Blood samples were collected from participants after an overnight fast into tubes appropriate for the collection of plasma (TAG. citrated plasma (PAI-1 and ICAM-1).5 5.068). HC/LGI vs control HS: P ¼ 0. while HDL cholesterol was slightly decreased after HC/HGI diet compared to the baseline (P ¼ 0.2±4.4) 9 (20) 5. Nutrient intake The mean daily intake of macronutrients is given in Table 2. In both low GI diets. CA.6±0. whereas it decreased in HC/HGI. HC/HGI vs HM/LGI: P ¼ 0.6 4. there was a small increase in protein intake in both high CHO diets compared to baseline (P ¼ 0. UK). Mo Circulating soluble ICAM-1 was measured by quantitative sandwich enzyme immunoassay (R&D Systems Inc..4 62. MINMOD Inc. some changes were observed when comparing the five interventions with the relative baseline.6± 28. Hampshire. Despite the diets being designed as iso-energetic. R&D Systems Europe Ltd. also reaching statistical significance in the HM/LGI group (Po0.7 1. but statistically significant.022) compared to baseline and between each diet (Po0.25 Glucose effectiveness (Sg) and insulin sensitivity (Si) were estimated with the MINMOD Millennium program (6.01).7±0.6 2 (4.1±9.0 137. Mainz. as starch increased in HC/LGI compared to baseline.01).0±0. The carrier gas helium was delivered at a flow rate of 14 ml minÀ1.0048). The participants were classified as being ‘at risk’ of developing the MetS when they scored X4 points.017.01). triacylglycerol.6 134. West Lothian. However. There were no changes in protein intake between diets.4 60.5±0. Cambridgeshire. Characteristics of study participants at screening M (n ¼ 43) 52. maintained for 0.3±0.

7±8.1±2.4 T 5.4±0.6±0.8** 11.1 5.0 132. T) with the five experimental diets (HS.8 642±193 558±137*wz 5.3 5.5±0.8 T 72.5* 79.1 4. for HC/HGI and HC/LGI) and HC/HGI also decreased plasma insulin concentrations (P ¼ 0.21±1.5±8.7±53.2±1. HC/HGI.5±4.4 50.63 32. HM/LGI.8±2.6 1.5±0. HGI.7 T 42.4 16.5±6.7 1.2 42.1 T 142.38 International Journal of Obesity (2013) 216 .2±1.5 5.7±2.7±6. B) and after 24 weeks dietary intervention (treatment.4±52. & 2013 Macmillan Publishers Limited 238.8±13. high carbohydrate.6±0. HC.4±0.9±1.5±3.9** 6.9±1.3 28.6±17.2 30.5±6.9±2.1 MUFA (%) B 12.5±9.3±4.3 89.7 82.1 HC/HGI (n ¼ 21) 5.78±2. while it decreased after intervention with HC/LGI (558±137 vs 642±193.9±0.9±36.8 19.6±0.5±5.1±3.2 T 17. mean±s.0 93.0±2.015 and P ¼ 0.4 LDL-cholesterol (mmol lÀ1) B 3.9** 1.5 1.3 1.7 CHO (%) B 43.5±6. carbohydrate.1 20.9 79.5** 42.6±0.8±2.3 110.1 27.8±34.2 TAG (mmol lÀ1) B 1.3±1.3±0.1±1. Table 3).1±3.7±3.0±36.6 14.1±32.1±7.1±12.7±6.1±5.6±93.0±3.3 Total cholesterol (mmol lÀ1) B 5.9±12.4±1.3 3. Comparisons between diets showed no significant effect of the dietary interventions on insulin sensitivity parameters calculated from the intravenous glucose tolerance test data (Si and Sg.8±80.8 NEFA (mmol lÀ1) B 558±134 T 621±202w Glucose (mmol lÀ1) B 5. HM/LGI.4±0.0±0.2 Protein (%) B 14.0±1.6 622±236 626±184 5.4 T 27.9±49.4±35.1 102.8±3.3±7.1±4.58±1.9±8.3 32.4 17.7±2.3±2.4* 29.5±48.microbiota interactions in obesity F Fava et al 219 Daily macronutrients intake estimated by analysis of 4 day food diaries collected from volunteers after the 4 weeks run-in diet (baseline.5 59.3 1.5 Starch (g) B 139.6** 15.3 1.9 Sugar (g) B 108.4±4.6 14.9 62.6 33.9 T 17.6±2.7±1.45±2.0±6.3 255.3±1.3±1.6 41.9 5.4 13.3 3. P ¼ 0.Diet .9 26.0 54.7±1.3 17.d.1±4.7±2.8±36. HC/LGI) HS (n ¼ 11) Energy (kcal) B 2153±443 T 2127±470 Fat (%) B 39.1±1.9±5.8 33. with HC/HGI compared to the baseline (P ¼ 0.8±9.01).97 235.8 92.7 16.4±3.1 17.8 15.4±1.5±1. PUFA.0 HM/LGI (n ¼ 22) 2005±578 2019±591 36.1±16. polyunsaturated fatty acid.6±0. MUFA. Values are expressed as mean±s.0 16.49±2. high glycemic index.2±65.1 19.3 T 101.0±25.4 682±242 573±144 5.5 19.7 70.8 3.3±30.5±6.4±4.1±17.5±0.0 259. LGI. mean±s.7±0.4* 251. Fasting plasma glucose concentrations were significantly lower after intervention with both HC diets compared to baseline (P ¼ 0.2 GI B T 63.7±8.5±0.8±0.) (Table 3). T) with the five experimental diets (HS.4 vs 238. biochemical data and IVGTT after the 4 weeks run-in diet (baseline.6 HDL-cholesterol (mmol lÀ1) B 1.8 1.6±12.5 5.9 3.9 65.7 17.028).8±8.1 89.3** 128.1 83.1 6.2±11.2 67.4 31.013).93±1.7±8. high monounsaturated fat.6±48.6 HC/HGI (n ¼ 21) 1920±569 1645±479* 35.7±2.17 3. HC/HGI.6 5.7±0.6 8.8±0.6* 73.3±8.0±1.3±2.4±9.2 17.7 17.6±2.48 1. mmlÀ1.0 16.7±5.6±44.6±58. HS: high saturated fatty acid.8±35. *.4 20.3* 62.2** 31.9±0.1±2.4±4.5 13.2** 16.7±1.1 61.9 PAI-1 (m mlÀ1) B 11.4±0.7±0.4±35.6 67.9 15. B) and after 24 weeks dietary intervention (treatment.7* 1.6±35.1±14.8 20.3 37.8±41.3** 5.67 T 3.8* Table 2.2 87.4 95.6±10.7 T 1.3 29.2±1.7±33.3±3. Table 3.8 54. Statistical conclusions did not change following adjustment for weight loss.9±1.2±0.8±6.4* 3.7±4.3±37.4 Leptin (ng mlÀ1) B 15.4±10. **Indicate significant differences with baseline values of the same diet (*Pp0.8±4.6** 65.5±9.5±7. Soluble ICAM-1 was higher after HM/HGI compared to baseline values (275.8±3.1 275.9±7.3 1.6±0.2±2. HC/LGI) HS (n ¼ 11) 5.4±0.7 88.0±0.5±0.8±59.4 107.6±0.0 T 17.7 2.6±3.2±6.9±4.4±37.6±48.1 103. low glycaemic index.9 78.0 18.045.6±2.2±56.6±0.1±3.3±39.8 HC/LGI (n ¼ 17) 5.7±37.3 136.6±0.3 17.2 HM/LGI (n ¼ 22) 5.9±5. Metabolic score at screening.4 ICAM-1 (m mlÀ1) B 232.0±24.8 1.0 27.4 158.3 93.7±5.7±0.7±0.5 32.0 2.8±21.7 1.3 259.3±2.3 3.8 PUFA (%) B 6.8 4.5 127±16 130±16 82±9 84±9.4±5.9±16.4±7.9±3.1±51.).4±0.5±7.05.6±7.6** 97.0 65.0** 11.8 T 3.8 5.1 5.4 T 5.8±14.7 137.3 92.3±13.8±2.4±0.6±3.6 45.9 CRP (mg lÀ1) B T 1.4 136±14 136±14 83±8 83±8 SBP (mm Hg) B 127±12 T 129±12 DBP (mm Hg) B T Body fat % B T 78±8 78±8 30.3 T 1.0±0.86±1.223 .75±1.7 75.5±3.1** 6.0±35.9±10.1±1.13±1.5* 1.0 T 16.6±8.1 T 37.9 T 11.02±1.4±11.5±0.0 16.9* 1.9 HM/HGI (n ¼ 17) 5.5±9.5** 143.9±2.d.7±6.9±10.3** 15.2 30.4±3.2±0.5 5.1±3. high monounsaturated fatty acid.9 64. HM/HGI.4 121.3±10.1±32.3 98.5 8.3 SFA (%) B 17.9* 1.8±0. **Pp0.5 28. HM/HGI.7±6.1±2.9 3.5 15.7±65.2** 44.0. non-starch polysaccharide.1 T 5.12 IS ( Â 10À4 ml mU-1 minÀ1) B 3.4±8.0 81.1±8.2±22.9±3.5±8.3 4.4 1.7 6. high saturated fat.3 Insulin (rmol lÀ1) B 77.3 2.9±1.2 19.8* 1.1 2.4±20.85 3.1 6.0 16. NSP.4±2. respectively.4±0.3±1.0±1. There were no significant changes in plasma TAG compared to baseline for any of the five diets.0 246.8** HC/LGI (n ¼ 17) MS 2212±426 1854±314** 37.8±59.28 1.7 538±129 617±165*z 5. HM.7 3.9±13.5±25.3** 11.9±45.3±3.5 16.2±1.3±2. P ¼ 0.4 3.8±29.6±52.5 129.4 WT (kg) B T WC (cm) B T 75.0±14.5±2.4 T 10.3 NSP (g) B 18.7 5.6±6. SFA.d.3±31.8±25.7±6.5±0.8 3.2±2.2±5.8 97.7±2.6 20.3±7.5 22.55 2. mmol lÀ1.4 9.5±0.5±0.4±0.3 Abbreviations: CHO.034.8 29.7* 1.9 66.0±18.3±13. anthropometric measurements.9±29.044.7 21.8±2.6 9.4 T 250.4±0.0±2.8* 1.0 127±14 126±19 79±11 78±12 BMI (Kg/m2) B 27.5 93.7±10.9 92.1 13.5±0.6* 5.9 9.21 3.2 80.4 29.5±0.6±2.7 55.2 68.3** 43.7** 12.8 104.1 14.9±7.34** HM/HGI (n ¼ 17) 2136±499 2056±418 39.5 35.1 18.7 128±13 127±14 78±7 78±11 34.3 1.6±8.6±5.8±1.7±0.2 96.9 6.9 5.1±1.6±0.

Sg.223 Figure 1.4±7 Abbreviations: BMI.005 and 9.3 vs 9. P ¼ 0. metabolic score at screening.d. The decrease in LDL-cholesterol was statistically significant for both low-fat/highCHO diets. SBP.0±3.9 17.4 17.2±4.005.009. respectively) (Figure 1).0022 and P ¼ 0.5 HM/HGI (n ¼ 17) 14. high monounsaturated fatty acid.9±0. insulin-independent glucose disposal.038).6±11. z Indicate significant differences between diets (Pp0. Numbers of Faecalibacterium prasnutzii increased after intervention with HS compared to baseline (9.7±2. although did not affect insulin sensitivity measures. though only the increase observed in the HC/HGI group was significant compared to control HS (P ¼ 0.45 (c).64 (b) and r ¼ À0.3 vs 9. propionate and n-butyrate).2 vs 10. high saturated fatty acid. mean±s. intravenous glucose tolerance test.7±0. mean±s.6±4. In contrast.4. insulin sensitivity. control diet (HS) compared to baseline (30. HT. P ¼ 0. HGI.2..7±0.6±0. but not compared to the other experimental diets.0 vs 6. for HC/HGI and HC/LGI).023 and 10. in this sub-group of & 2013 Macmillan Publishers Limited . r ¼ À0.5±0.05. height. mean±s.659 and 10. systolic blood pressure.2 vs 10.022). although this change was not significant in comparison with other diets. respectively. WT. P ¼ 0. HS.5. 7. No differences in SCFA levels were observed when comparing the five intervention diets. These results are in accordance with those reported previously for the entire RISCK study population (n ¼ 640) and confirm dietary compliance. BMI and waist circumference (Pearson’s correlation: r ¼ À0.3±3.d.25 Replacement of dietary fat for carbohydrate had a positive impact on fasting glycemia (HC/HGI and HC/LGI) and decreased fasted insulin levels (HC/HGI) compared to baseline. high carbohydrate.6±0. Log10 [bacterial cells per g faecal contents].. the number of total bacteria remained unchanged in the two diets with a lower fat content (10.05). respectively. Faecal SCFA concentrations The result of faecal SCFA concentrations analysis are shown in Table 5. P ¼ 0. r ¼ À0..2. w.0214).7±0.microbiota interactions in obesity F Fava et al 220 Table 3 (Continued ) HS (n ¼ 11) Sg ( Â 10À3 per min) B 16.7±0.9.d.d. for HM/HGI and HM/LGI) and in comparison with baseline values (10. P ¼ 0. Volunteers on HS also showed some decrease in total bacteria levels at the end of the study compared to baseline (10. P ¼ 0. HC. P ¼ 0. mean±s.7±0. mean±s.2. P ¼ 0. International Journal of Obesity (2013) 216 -. Total bacteria decreased after intervention in the three diets with the highest fat content.005.2 Log10[bacterial cells per g faecal contents].. However.8±3. MS.64. P ¼ 0.052.6 T 15.2 vs 10. for HC/HGI and HC/LGI).018) and HC/LGI (9. **Indicate significant differences compared to baseline values of the same diet (*Pp0.3.5±0.0148.d. r ¼ À0. Total bacterial numbers of both high MUFA.3 Log10[bacterial cells per g faecal contents]. mean±s. **Pp0.011.9±5 HC/HGI (n ¼ 21) 17.105).diet with the lower GI. mean±s. while it was significant only in the high MUFA..d.6 15. high glycemic index. DBS.(HM/LGI). However.8±0.4 vs 9.6±0. respectively. P ¼ 0.8±0.863.4±6. compared to baseline (9.9 vs 25. respectively. IVGTT.2. BMI (b) and waist circumference (c) in the HC/HGI intervention group.45. Moreover.d.3 Log10[bacterial cells per g faecal contents]. participants on the HC/HGI diet also had significantly higher levels of Bacteroides spp.6±0.2.01).d.6±3. HM.3 vs 10. the two HC diets showed the largest increase in Bifidobacterium spp.6±6. faecal numbers and body weight (a).6 vs 9.diets were significantly lower than HC/HGI (P ¼ 0.2±7. population levels compared to baseline values (9.. for acetate. Significant correlation between changes in Bacteroides spp. This increase in bacteroides numbers after HC/HGI diet was associated with decreases in body weight..2±0.8. weight.5 HC/LGI (n ¼ 17) 15. P ¼ 0. Log10[bacterial cells per g faecal contents].6 vs 8. mmol lÀ1. Log10[bacterial cells per g faecal contents].5±0.2±10. Microbial enumeration in faecal samples by FISH Significant changes were observed in total bacteria number at the end of the study compared to baseline in some of the experimental groups (Table 4).7 19. diastolic blood pressure. respectively). the fecal concentrations of acetate.0±0. waist circumference. mean±s. low glycaemic index. DISCUSSION Dietary intervention with both high-MUFA diets and with both low-fat/high-CHO diets decreased plasma total and LDL-cholesterol concentrations in comparison with baseline.6±0. for HM/HGI and HM/LGI.64 (a). propionate and n-butyrate all increased after the high saturated fat. IS. P ¼ 0. WC.7±0.8±4. body mass index. regardless of the GI.2±0. Interestingly. LGI.3 HM/LGI (n ¼ 22) 17±4.Diet .64. 7. r ¼ À0. Pearson’s correlation.. Log10[bacterial cells per g faecal contents]. *. P ¼ 0.2 vs 10.8 vs 6.

Moreover.04 34.8±0.84±0.5 9.5 8. T) with the five experimental diets (HS.0±0.49 33.73±0.3 9.5±0.7±0.79** 7.4 8.6±0. and their growth may have been stimulated by the increased bioavailability & 2013 Macmillan Publishers Limited of dietary carbohydrate. A lower prevalence of Bacteroidetes has been reported for both animal models of obesity (ob/ob mice) and in obese humans compared with their lean counterparts.43 7. which to a diet include pectin.3* 8. B) and after 24 weeks dietary intervention (treatment.20±1.4 Acetate B T Propionate B T i-Butyrate B T n-Butyrate B T i-Valerate B T n-Valerate B T n-Caproate B T Table 4.3 9.16 However. **Indicate significant differences with baseline values of the same diet (*Pp0. 40 -6-diamidino-2-phenylindole.4* 9.7±0.46 0. low glycaemic index.2 10.5±0.49 0.2 10.14±10. Nevertheless.0±0. HS: high saturated fatty acid. HM/LGI.00 9. inulin.00±0.3 8.6 HM/HGI (n ¼ 17) 10.98±10.01).40 1.3 8.3 9.82±0.05). high monounsaturated fatty acid.74 0.2±0. the fibre content).56±0.2±0.2±0.4 10. T) with the five experimental diets (HS.42 6.47 1.4 8.84 7. Other complex carbohydrate common in the human ˆ -glucan.2±0.7±0.3wz 10.3±0.223 .81 0.42±0. high glycemic index.5 8.55±0.microbiota interactions in obesity F Fava et al Microbial enumeration by FISH in faecal samples collected after the 4 weeks run-in diet (baseline. one measure of dietary fibre.5 8.6** 8. **Pp0.34 0.8±0.54 0.1±0.46 1.d.58±0. HGI.4±0.6 8. HC/LGI) 221 Table 5.44 0.6±0.0±0.89±0.3 9.45±6.75 0.2±0.3 9.9±0.94±1. In the RISCK study.43 0.86±0.05.2*w 10.84±3.43 Cani et al.0±0.3* 9. significant insulin and percentage of body fat decreases were observed relative to baseline.60 9. high monounsaturated fatty acid.51±0.46 0.02±3.7±0.69 0.21±12.6 8.7w 8.11 1.74 0.6±0. LGI. was significantly higher in HM/LGI but not in HC/LGI compared to baseline.67 0.99±0.). **Indicate significant differences with baseline values of the same diet (*Pp0.6±0. in those participants with a significant and marked faecal bifidobacteria increase (HC/HGI).31 0. waist circumference and BMI (Figure 1). A significant decrease in fasted blood glucose was observed after both high-carbohydrate/low-fat diets compared to baseline.31±3.65±0. *.34±0.56 32.7 8.60±4.13±3.16.5±0.4±0. mean±s.34 0. high saturated fatty acid.49±0. B) and after 24 weeks dietary intervention (treatment.3±0. HM/HGI.5±0.4 9.8±0. HC/LGI) Probe/DNA stain DAPI B T EREC482 B T Bac303 B T Ato291 B T Fpra655 B T Bif164 B T Lab158 B T Chis150 B T HS (n ¼ 11) 10.3 9.12.76±0.8 8.4 HM/LGI (n ¼ 22) 10.5±0.5 9.7±0.43 0.7±0.4 8.67 0.7±0.1±0.4 8.4±0.7±0.99 9.63 1.65 0.5±0.5±0.3 9.3 8.72±3. measured by gas chromatography after the 4 weeks run-in diet (baseline. although the resistant starch intake or percentage of fermentable fibre were undetermined. the glycemic response) and not an inherent chemical property of a food (that is.48 0.1±0.4 9.57±14.4 8.38±4. mean±s.85±0.4 9.78 0. Participants on both low-fat high-CHO diets also had significant increase in Bifidobacterium and showed a modest increase in Atopobium numbers. high carbohydrate.35 0.4 9. These dominant members of the human gastrointestinal microbiota are important degraders of carbohydrate. Interestingly. Abbreviations: DAPI.4±0. HM/LGI.17 8. LGI.7±0.88±0.15 31.3 8.69 6.3 9.5±0.7 8.36±12. Reducing dietary fat intake and increasing dietary carbohydrate consumption increased both faecal Bacteroides and Bifidobacterium spp. Bacterial numbers are expressed as log10 [cells per g of faecal content wet weight].60 0.2 10.3 9.72 9. short chain fatty acid. *.5 9.57±3.4±0.5±0.3 9. We investigated the long-term impact of different amounts and quality of dietary fat and carbohydrate on gut microbial composition and fermentation activities in a population ‘at MetS risk’.7±0.5 9. groups of bacteria that have been independently linked to improved body energy regulation and reduced risk factors of MetS.0±0.4 9. the diets were designed to maintain body weight. HS.3 10.5±0.21±10.65 0.61±11.d.30 29. SCFA.59 1.8±0.5±0.52±0.2 10. both within the Actinobacteria phylum.3 10.75 0. this increase was not significant and from the data analysis only occurred in a minority of individuals.98±0.4 9.2±0.09 7.22 showed an association between the bifidogenesis following dietary supplementation with oligofructose and improved MetS International Journal of Obesity (2013) 216 .2±0.5 8. HC.5±0.95 0.05±0.4±0.7±0. HS (n ¼ 11) HM/HGI (n ¼ 17) HM/LGI (n ¼ 22) HC/HGI (n ¼ 21) HC/LGI (n ¼ 17) 25. HGI. HC/HGI. there was a concomitant increase in Bacteroides.84 Abbreviations: HC. it should be considered that GI is a measure of a physiological response to a food (that is.80 7.93* 31.79±0.3 8.3 9.42 and can be augmented with prebiotic fibres ingestion.9±0.90 7.62 8.05. low glycaemic index. RISCK study participants (n ¼ 88) no changes between diets were observed.76±4.7±0.7±0.2 10.2 9.2±0. In the present study.77±0. there were no significant differences in the total starch or non-starch polysaccharide intakes.6 HC/LGI (n ¼ 17) 10.60 0.8±0.3 9.2 9.8±0.39 0..84±5.04±13.3 9.8 9.3 10.8±0. when a small loss of weight was observed.9 8.6*w 8.52±8.88±0. no details were given on dietary design. Bifidobacteria are beneficial members of the gastrointestinal microbiota.76±0.6 HC/HGI (n ¼ 21) 10.15 Weight loss following 1-year fat/CHO-restricted diets was shown to correlate with increased Bacteroidetes in humans.71±0.3 10.2±0.3**z 10.1±0.6±0.2 10.8±0.3 Log10[bacterial cells per g faecal contents] in the HM/HGI diet group. Studies in ileostomy patients have shown that the amount of starch reaching the colon is directly proportional to the quantity of starch ingested.7 8.5 8.6±0.69 0.7±0.28±0.1±0.3 9.22 Increased Bacteroides numbers after the HC/ HGI diet directly and significantly correlated with a modest decrease in body weight.70±0.90±0. **Pp0. Faecal SCFA (mmol lÀ1.71 0.3 8.2±0. HM. HM/HGI.28±3.92±5.19±4.3±0. Moreover. Non-starch polysaccharide.79 7. w. high glycemic index.2 9.61 8.21±6.77 0. high carbohydrate.7±0.96** 7.70 30.25 0.2 10.79±20. HC/HGI. arabinoxylan and a large degree escape digestion in the upper gut and reach the colon.01±0. z Indicate significant differences between diets (Pp0.89±0. a trend towards decreased body weight and BMI was observed after HC/HGI and HC/LGI. HM. Although mean counts of faecal bifidobacteria increased by 0.2±0.3 9.6±0.87±4.41 This is important.12±4.16±0.Diet .6±0.77 1.9±0. where they are readily fermented.44±0.01). as starch is the major constituent of most human carbohydraterich foods.92±0.3 9. While it might be expected that low-GI diets are enriched in fibres.0±0.79 1.97±14.56±0.2±0.61±0.2±0.72±0.60 7. a trend towards decreased body weight and BMI was observed after HC/HGI and HC/LGI.17 31.32±3.87±7.6±0.79 32.40.99 8.32 7.4 9.04±0.84 0.

1023. Beijer S. Nature 2011.Diet . ¨ tta ¨la ¨ R. We report a significant association between faecal bacteroides and weight loss and present data to show that HC diets stimulate faecal bifidobacteria. obesity and time in mouse models. Frost GS. O’Sullivan O. Holmes E. Belenguer A. 103: 12511 . Turnbaugh PJ. An obesity-associated gut microbiome with increased capacity for energy harvest. 444: 1022 . Stern MP.758. Lozupone CA. Fava F. Iglesias MA. physical activity and diet: England. Do non-insulin-dependent diabetes mellitus and cardiovascular disease share common antecedents? Ann Intern Med 1996. Gitau R.195. Changes in gut microbiota control inflammation in obese mice through a mechanism involving GLP-2-driven improvement of gut permeability. Selective increases of bifidobacteria in gut microflora improve high-fat-diet-induced diabetes in mice through a mechanism associated with endotoxaemia. Arterioscler Thromb Vasc Biol 1995. Top-down systems biology modeling of host metabotype-microbiome associations in obese rodents. Healy S. Moore C. Gordon JI. Griffin BA. Comparative effects of very low-carbohydrate. Conterno L. as acetate. Rachel Gitau recruited the volunteers. Lovegrove JA. Charles MA. Gardner CD. Knight RD. 6: 241 . Levison BS. ¨fer K. Francesca Fava helped with volunteer recruitment and IVGTTs. Borch-Johnsen K. Monounsaturated versus polyunsaturated dietary fat and serum lipids. Burcelin RG.44 Higher SCFA observed in our HS group might reflect lower absorption. Ley RE. 18: 190 . Blancher C.15 despite a lower prevalence of the major polysaccharide degrading gut bacteria. and an 25 26 alternative definition of an insulin resistance syndrome. Effect of changing the amount and type of fat and carbohydrate on insulin sensitivity and cardiovascular risk: the RISCK (Reading. Kraemer HC. 37: 1595 . Amar J. Everard A. Ba Obesity alters gut microbial ecology. a finding that may implicate an aberrant production or absorption of SCFA in response to a modern diet in people at increased metabolic risk. fat mass development and inflammatory tone) in high-fat fed mice.11075. Mart´ Balca ınez-Abundis E.116. Clohessy AM. Successful manipulation of the quality and quantity of fat and carbohydrate International Journal of Obesity (2013) 216 -. Gordon JI. Proc Natl Acad Sci USA 2005. Tuohy KM. Wilson ID. Jebb SA. Metabolic profiling reveals a contribution of gut microbiota to fatty liver phenotype in insulin-resistant mice. 131: 597 . performed faecal microbiota analysis and wrote the manuscript. Ley RE. Jackson KG. Proc Natl Acad Sci USA 2006.pdf. Turnbaugh PJ. Ley RE. University of Surrey (Bruce Griffin).223 & 2013 Macmillan Publishers Limited . 92: 748 . Duncan SH. Marques TM. Turnbaugh P. Lifestyles Statistics (10 February 2010). Flint HJ. 59: 1635 .1607. Williams CM. Chatfield MD et al.2383.300. Knauf C. MRC Unit of Human Nutrition Research. 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 CONFLICT OF INTEREST The authors declare no conflict of interest.1772. Diet-induced obesity is linked Turnbaugh PJ. Effect of oral inulin ˜ oz BR. Viola R. Cloarec O. In conclusion.12516. 26: 5 . Higher faecal SCFA levels might be due to decreased intestinal absorption or a modulation of the gut microbiota activity or composition (for example. 11: 294 . Gordon JI. The effect of the daily intake of inulin on fasting lipid. Clifton PM. Wang Y. Drivsholm T. Holtrop G. Taylor GR. Ba to marked but reversible alterations in the mouse distal gut opad10/Statistics_on_Obesity_Physical_Activity_and_Diet_England_2010. 19 20 21 22 23 24 REFERENCES 1 The NHS Information Centre. Cotter PD. Rantalainen M. Bennett BJ. ¨ ckhed F. Cambridge. Obesity and the human microbiome. Bird AR. a decreased bacterial populations that utilise these fermentation products as a source of energy). Nature 2006. Mardis ER. UK).63. Reduced dietary intake of carbohydrates by obese subjects results in decreased concentrations of butyrate and butyrate-producing bacteria in feces.2375.1103. 444: 1027 . Dumas ME. ´zar-Mun ´ lez-Ortiz M. 56: 1761 . Klein S. Lobley GE. Br J Nutr 1999. Appl Environ Microbiol 2007. Magrini V. Microbial ecology: human gut microbes associated with obesity.223. Scha Microbiota and SCFA in lean and overweight healthy subjects. Surrey. Klipfell E. Br J Nutr 2009. Cereal Partners UK (Welwyn Garden City.260. Bastelica D et al.19 also reported higher SCFA in obese humans compared with lean. Weetabix Ltd (Kettering. Gordon JI. this study provides evidence that both the type and quantity of dietary carbohydrate and fat impact on gut microbial composition and activity in people at MetS risk. Laaksonen M.604. rather than higher colonic fermentation. Obesity and the gut microbiota: does up-regulating colonic fermentation protect against obesity and metabolic disease? Genes Nutr 2011. 50: 2374 . Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Role of insulin resistance in human disease. Pra Eur J Public Health 2001. UK). Guiot Y.43 Vogt and Wolever demonstrated that the faecal acetate percentage inversely correlates with the absorbed acetate percentage after rectal infusion. ¨ckhed F. insulin and glucose concentrations in middle-aged men and women. Metabolic endotoxemia initiates obesity and insulin resistance. Koeth R. Gonza administration on lipid profile and insulin sensitivity in subjects with obesity and dyslipidemia. Rottier O et al.376. Obesity 2010. Nature 2006. Donus C et al. collected faecal samples. 101: 1493 . Rev Med Chil 2003. Netherlands). A meta-analysis. Taras D. body weight gain. HM/LGI) in this study were accompanied by decreased total bacteria compared to baseline. Cani PD. Schwiertz A. Poggi M. Mahowald MA. 58: 1091 . propionate and butyrate are mainly microbial saccharolytic end products. Cani PD. Composition and energy harvesting capacity of the gut microbiota: relationship to diet.22 Despite the observed changes in microbial population levels. Wang Z. an established gut health biomarker. Brinkworth GD. UK) and Sainsbury’s Supermarkets Ltd (London. Bos NA. Curr Opin Gastroenterol 2010. Reaven GM. We also observed increased faecal SCFA after a HS. without a concomitant decrease in any of the FISH-enumerated bacteria. Waldram A.30. Gut 2009. run the dietary intervention and performed IVGTTs. Griffin MD. Knauf C. 3: 213 . Schwiertz et al. Chatfield MD et al.1078. Fulton L. 2 Balkau B. HM/HGI. The high fat diets (HS. Yudkin JS et al. Available at http://www. Hertfordshire. Barton RH. Diabetes 2007. Statistics on obesity.microbiota interactions in obesity F Fava et al 222 biomarkers (improved glucose tolerance. 18 ACKNOWLEDGEMENTS We acknowledge the contribution of each RISCK study center member: University of Reading (Julie Lovegrove). high-fat and high-carbohydrate. Carmel Moore) and Kings College London (Tom Sanders).11. Ley RE. low-fat weight-loss diets on bowel habit and faecal short-chain fatty acids and bacterial populations. insulin sensitivity. Goff L. suggesting that the employed FISH probes did not cover all the faecal bacterial populations and that a broader spectrum technique (that is. Tuohy KM et al. Possemiers S. 8: 2361 . Foods were supplied by Unilever Food and Health Research Institute (Unilever R&D. University of Cambridge (Susan Jebb. Wareham N. Noakes M. 15: 1917 . Dugar B et al. 73: 1073 . 28: 364 . Imperial College London (Gary Frost). Johnstone AM. Recent studies have shown that caecal SCFA levels are higher in obese animals compared with lean. Imperial. Diabetes 1988.1927. These findings were unexpected. Frequency of the WHO metabolic syndrome in European cohorts. Cell Host Microbe 2008. no significant changes were observed in faecal SCFA after any of the experimental diets except the HS control diet. high throughput sequencing) might be more appropriate to monitor subtle gut microbial changes. The Health and Social Care Information Centre. J Proteome Res 2009. 124: 110 . Moore CS.1502. Am J Clin Nutr 2010. This study was supported by the UK Food Standards Agency (project NO2031). Rothwell A et al. Cani PD. 102: 11070 . Neyrinck AM. Toye A. 472: 57 . Banting lecture 1988. Diabetes Metab 2002. Kieran M Tuohy and Glenn R Gibson coordinated the faecal microbiota analysis and helped with writing of the manuscript. Katie Newens helped with volunteer recruitment and IVGTTs.nhs. Fouhy F et al. Gut 2010. Tuohy KM et al.ic. In particular for the University of Reading: Julie A Lovegrove was the principal investigator for the Reading Centre. and Kings) trial. 2010. Vlaardingen. Grampian (Banff. Diabetologia 2007. Murphy EF.1031. Lewis FJ. UK). Van de Wiele T. Patterns of unhealthy behaviour in Finland. Fava F. Karisto A.1642. 82: 23 .

Harmsen HJ. Gastrointestinal microflora: probiotics. Gordon JI. Variations of bacterial populations in human feces measured by fluorescent in situ hybridization with group-specific 16S rRNA-targeted oligonucleotide probes. Jansen GJ. Manz W. Bru ¨ ck WM. Science 2005. Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bacteria of the phylum cytophaga-flavobacter-bacteroides in the natural environment. 69: 4320 .12. Jansen GJ. Appl Environ Microbiol 1998. Microb Ecol Health Dis 1999. 61: 3069 . J Nutr 2009.223 . Jansen GJ. 307: 1915 59: 187 . 64: 3336 . Degener JE.3345. Knol J. Ootsubo M. Obes Rev 2007. Schut F. 8 (Suppl 1): 93 . Saulnier DM. Grijpstra Raangs GC. Tajima K. Gibson GR.3148. ¨ ckhed F. 93: 60 . Harmsen veys/ndnsdocuments/. J Appl Microbiol 2002. Saunders DR. Yoshimizu M et al. 139: 1534 . Graham MM.3345. 53: 43 . Modulation of the human gut microflora towards improved health using prebiotics--assessment of Ley RE. Peterson DA. 86: 341 . Host-bacterial Ba mutualism in the human intestine. Tuohy KM. 36 29 30 37 38 31 39 32 40 41 33 42 43 34 35 44 & 2013 Macmillan Publishers Limited International Journal of Obesity (2013) 216 . Harrison ST. 64: 3336 .97. 11: 75 . Oligonucleotide probe for detecting Enterobacteriaceae by in situ hybridization. Variations of bacterial populations in human feces measured by fluorescent in situ hybridization with group-specific 16S rRNA-targeted oligonucleotide probes. 41: 1244 .4324. Schut F. Schut Gibson GR. with genus-specific 16S rRNA-targeted probes and its application in fecal samples. 1997-2000. 66: 4523 . Harmsen HJM.68. A 16S rRNA-targeted probe for detection of lactobacilli and enterococci in fecal samples by fluorescent in situ hybridization.90. Sillery JK. Appl Environ Microbiol 1998. Moosa S. A study of the relative dominance of selected anaerobic sulfate-reducing bacteria in a continuous bioreactor by fluorescence in situ Development of 16S rRNA-based probes for the Coriobacterium group and the Atopobium cluster and their application for enumeration of Coriobacteriaceae in human feces from volunteers of different age groups. Icgen B. Quantitative fluorescence in situ hybridization of Bifidobacterium spp.1106. Sawabe T. Available from: http://tna. Schut F. Schleifer KH. Dietary determinants of obesity. Sonnenburg JL. Tanaka R.3075.1248. Wildeboer-Veloo Kamphuis GR. Shimizu T. Adv Appl Microbiol 2006. Franks AH. Oligonucleotide probes that detect quantitatively significant groups of butyrate-producing bacteria in human feces. National Food Survey (NFS). Kolida S. Vancanneyt M. Schwiertz A. Appl Environ Microbiol 1995. Elfferich P. Gibson GR.52. Raangs GC. Lustenberger AM. http://www. Absorption of starch by healthy ileostomates: effect of transit time and of carbohydrate load. Harmsen HJ. Hold GL. Welling GW. 11: 3 . Kolida S. National Diet and Nutrition Survey (NDNS) 2004. 133: 3145 . Wilkinson MH et al. Available from: http://tna.1540. Microb Ecol 2007. Langendijk PS. Am J Clin Nutr 1985. Blaut M. Curr Pharm Des 2005. Microbiology 1996. Raangs GC. J Nutr 2003. Tuohy KM. Flint HJ. Franks AH. 142: 1097 . Welling GW. Fecal acetate is inversely related to acetate absorption from the human rectum and distal colon. Vogt JA.4527. Amann R.Diet .microbiota interactions in obesity F Fava et al 223 27 28 consumed by free-living individuals using a food exchange model. Appl Environ Microbiol 2000. europarchive. Welling GW. The prebiotic effects of biscuits containing partially hydrolysed guar gum and fructo-oligosaccharides--a human volunteer study. Aminov RI. Appl Environ Microbiol 2003. Rouzaud GC. Welling GW.219. Br J Nutr 2001. Wolever TM. http://www. Ludwig W. Chapman RW. Jebb SA.

Sign up to vote on this title
UsefulNot useful